Hierarchical Techniques to Forecast Joint Price-Volume Distributions of Intraday, Continuous Power Markets

Intraday electricity markets (IDM) facilitate continuous buying and selling of power after the closure of the day-ahead market (DAM) for the same and the following day.Each delivery hour in these IDM markets experiences thousands of transactions, resulting in a range of prices for distinct volumes due to the pay-as-bid pricing model.Modelling the joint price-volume (P-V) behaviour is beneficial for power traders, energy companies, and consumers to design optimized buy and sell bids.But, predicting both price and volume together poses significant challenges due to the complex and dynamic interactions influenced by market regulations, volatility in renewable energy generation and consumer behaviour.This paper proposes a hierarchical approach to model price-volume (P-V) distribution conditioned on the traded time.The first stage involves learning latent representations of the P-V distributions using a CNN-based autoencoder, while the second stage leverages these learnt latent representations, along with exogenous variables such as weather and calendar data, to train an LSTM model.The resultant LSTM outputs are then decoded back to the original distributions using the CNN autoencoder.The proposed methodology is trained and tested on EPEX Intraday market data.When benchmarked against intuitive baselines, the proposed approach significantly outperformed the baseline method.